Make the asset-version guard line-ending independent

solvreven commented

2026-09-02 20:42:00 +02:00

Owner

Outcome

test_asset_version.py hashed the four versioned assets byte-for-byte. They are text files, the repo has no .gitattributes, and CI checks them out LF on debian-trixie-python while a Windows clone with core.autocrlf=true holds CRLF. The digest therefore differed by platform, and the guard was green only for whoever recorded it last.

Both sides have been re-recording their own platform's value in turn, each fix correct locally and each breaking the other:

commit	recorded	LF (CI)	CRLF (Windows)	CI	Windows
`1c94db7` v135	`d20d1fa11d2d9a94`	`a932c205c61efb9b`	`d20d1fa11d2d9a94`	RED	green
`bbf3c01` v136 (PR #27)	`a932c205c61efb9b`	`a932c205c61efb9b`	`d20d1fa11d2d9a94`	green	RED
`9c2faca` v137	`d9ca43b746e10252`	`350840c64b1a1aa0`	`d9ca43b746e10252`	RED	green

So main has been red in CI since 9c2faca, and PR #27's asset commit was not wrong — it was right for Linux.

Why this shape of fix

The repo had already solved this, one digest over. projection_digest in pipeline.py normalises \r\n to \n and says why, in terms that describe this bug exactly: "git stash/checkout rewrite tracked files between LF and CRLF under core.autocrlf=true, and on Windows that flipped this digest with zero code difference: same source, two different hashes, one commit apart." It is covered by test_crlf_and_lf_produce_the_same_digest.

The asset digest is the one place that remedy was never applied — "fixing a defect in one location is not fixing the defect", with the second location in a different file.

Chosen over a .gitattributes entry deliberately: that would leave the guard's correctness depending on every contributor's checkout config, which is the disconnected-mechanism shape this repo keeps paying for. Normalising in the digest makes the platform difference unrepresentable instead.

A CRLF/LF flip is not an asset change a browser needs a new ?v= for, so it must not move the digest. EXPECTED_VERSION stays 137; EXPECTED_DIGEST becomes the normalised value, identical under both conventions.

Verification

Proved RED on both the property fixed and the property that must not break — driving the real test module against controlled inputs:

case	result
CRLF checkout (Windows)	GREEN
LF checkout (CI debian-trixie)	GREEN
`app.js` changed, `?v=` not bumped	RED
`tokens.css` changed, `?v=` not bumped	RED

tests/web/test_asset_version.py: 3 passed. Nothing else in the tree references that module or its _digest, so the change is self-contained.

Note for #27

bbf3c01 becomes unnecessary rather than conflicting once this lands. Asked there for a rebase, so neither side has to concede a digest that was correct where it was measured.

🤖 Generated with Claude Code

https://claude.ai/code/session_014adnmRztkdAPwinWXnNQ7N

## Outcome `test_asset_version.py` hashed the four versioned assets byte-for-byte. They are text files, the repo has no `.gitattributes`, and CI checks them out LF on `debian-trixie-python` while a Windows clone with `core.autocrlf=true` holds CRLF. The digest therefore differed by platform, and the guard was green only for whoever recorded it last. Both sides have been re-recording their own platform's value in turn, each fix correct locally and each breaking the other: | commit | recorded | LF (CI) | CRLF (Windows) | CI | Windows | |---|---|---|---|---|---| | `1c94db7` v135 | `d20d1fa11d2d9a94` | `a932c205c61efb9b` | `d20d1fa11d2d9a94` | **RED** | green | | `bbf3c01` v136 (PR #27) | `a932c205c61efb9b` | `a932c205c61efb9b` | `d20d1fa11d2d9a94` | green | **RED** | | `9c2faca` v137 | `d9ca43b746e10252` | `350840c64b1a1aa0` | `d9ca43b746e10252` | **RED** | green | So **`main` has been red in CI since `9c2faca`**, and PR #27's asset commit was not wrong — it was right for Linux. ## Why this shape of fix **The repo had already solved this, one digest over.** `projection_digest` in `pipeline.py` normalises `\r\n` to `\n` and says why, in terms that describe this bug exactly: *"`git stash`/`checkout` rewrite tracked files between LF and CRLF under `core.autocrlf=true`, and on Windows that flipped this digest with zero code difference: same source, two different hashes, one commit apart."* It is covered by `test_crlf_and_lf_produce_the_same_digest`. The asset digest is the one place that remedy was never applied — "fixing a defect in one location is not fixing the defect", with the second location in a different file. Chosen over a `.gitattributes` entry deliberately: that would leave the guard's correctness depending on every contributor's checkout config, which is the disconnected-mechanism shape this repo keeps paying for. Normalising in the digest makes the platform difference unrepresentable instead. A CRLF/LF flip is not an asset change a browser needs a new `?v=` for, so it must not move the digest. `EXPECTED_VERSION` stays 137; `EXPECTED_DIGEST` becomes the normalised value, identical under both conventions. ## Verification Proved RED on both the property fixed and the property that must not break — driving the real test module against controlled inputs: | case | result | |---|---| | CRLF checkout (Windows) | GREEN | | LF checkout (CI debian-trixie) | GREEN | | `app.js` changed, `?v=` not bumped | **RED** | | `tokens.css` changed, `?v=` not bumped | **RED** | `tests/web/test_asset_version.py`: 3 passed. Nothing else in the tree references that module or its `_digest`, so the change is self-contained. ## Note for #27 `bbf3c01` becomes unnecessary rather than conflicting once this lands. Asked there for a rebase, so neither side has to concede a digest that was correct where it was measured. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_014adnmRztkdAPwinWXnNQ7N

solvreven added 1 commit

2026-09-02 20:42:00 +02:00

fix: the asset guard hashed raw bytes, so it could only be green on one platform

CI / syntax (pull_request) Successful in 8s

Details

CI / tests (pull_request) Successful in 1m7s

Details

CI / wheel (pull_request) Successful in 29s

Details

CI / browser (pull_request) Successful in 16m33s

Details

83111ce2e0

test_asset_version.py hashed the four assets byte-for-byte. These are text
files and the repo has no .gitattributes, so CI checks them out LF on
debian-trixie while a Windows clone with core.autocrlf=true holds CRLF.
The digest therefore differed by platform, and the guard was green only
for whoever recorded it last.

Both sides then re-recorded their own value in turn, each fix correct
locally and each breaking the other:

  1c94db7  135  d20d1fa11d2d9a94  CRLF   green here, RED in CI
  bbf3c01  136  a932c205c61efb9b  LF     green in CI, RED here   (PR #27)
  9c2faca  137  d9ca43b746e10252  CRLF   green here, RED in CI

So main has been red in CI since 9c2faca, and PR #27's asset commit was
not wrong -- it was right for Linux. Neither side could see the other's
failure, because each measured the assets through its own checkout.

A CRLF/LF difference is not an asset change a browser needs a new ?v=
for, so it must not move the digest. Normalising before hashing makes the
platform difference unrepresentable rather than documented; recorded
digest is now the normalised value, identical under both conventions.

Chosen over a .gitattributes entry deliberately: that would leave the
guard's correctness depending on every contributor's checkout config,
which is the disconnected-mechanism shape this repo keeps paying for.

Proved RED on both properties -- the one fixed and the one that must not
break:

  CRLF checkout (Windows)              GREEN
  LF checkout (CI debian-trixie)       GREEN
  app.js changed, ?v= not bumped       RED
  tokens.css changed, ?v= not bumped   RED

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014adnmRztkdAPwinWXnNQ7N

solvreven referenced this pull request

2026-09-02 20:42:16 +02:00

Fix the asset guard and establish the odds router boundary #27

solvreven referenced this pull request

2026-09-02 20:44:31 +02:00

Norsk Tipping is the only book on the board, and it fetches itself #29

solvreven referenced this pull request

2026-09-02 22:29:40 +02:00

Understat: current-season npxG, a live squad list, and a chore that fetches it #30

solvreven referenced this pull request

2026-09-02 23:52:50 +02:00

Squad rail becomes a league-wide captaincy board, star-rated #31

CI / syntax (pull_request) Successful in 8s

Details

CI / tests (pull_request) Successful in 1m7s

Details

CI / wheel (pull_request) Successful in 29s

Details

CI / browser (pull_request) Successful in 16m33s

Details

This pull request can be merged automatically.

You are not authorized to merge this pull request.

View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.

git fetch -u origin fix/asset-guard-line-endings:fix/asset-guard-line-endings

git switch fix/asset-guard-line-endings

Merge

Merge the changes and update on Forgejo.

Warning: The "Autodetect manual merge" setting is not enabled for this repository, you will have to mark this pull request as manually merged afterwards.

git switch main

git merge --no-ff fix/asset-guard-line-endings

git switch fix/asset-guard-line-endings

git rebase main

git switch main

git merge --ff-only fix/asset-guard-line-endings

git switch fix/asset-guard-line-endings

git rebase main

git switch main

git merge --no-ff fix/asset-guard-line-endings

git switch main

git merge --squash fix/asset-guard-line-endings

git switch main

git merge --ff-only fix/asset-guard-line-endings

git switch main

git merge fix/asset-guard-line-endings

git push origin main

Rows
Columns

Make the asset-version guard line-ending independent #28

Outcome

Why this shape of fix

Verification

Note for #27

Checkout

Merge